Sorting LiteLLM GitHub issues by difficulty to implement. Higher = harder.

{
The right item requires a cross-cutting integration with runtime configuration, model execution, proxy and SDK pathways, observability, and fail-open behavior. The left item is a localized schema-processing correction with focused regression tests.
}
https://github.com/berriai/litellm/issues/29320 8:1 https://github.com/berriai/litellm/issues/27853